Can anyone clarify the correct backoff strategy when hitting rate limits during bulk user attribute updates? I am using the Python SDK to iterate over a list of users and call put_users_user(user_id, body). The documentation states: “Rate limits are enforced per tenant and per endpoint. Clients should implement exponential backoff.” However, my current implementation using a fixed 5-second delay still triggers 429 Too Many Requests errors after approximately 50 requests per minute.
I have included a simplified snippet of my current approach below. The Retry decorator from tenacity is configured with a fixed wait, but it seems insufficient for the Genesys Cloud API limits. Is there a specific header I should be parsing from the 429 response, such as Retry-After, to dynamically adjust the wait time? Or is there a recommended pattern for batching these requests to stay within the documented limits without manual tuning?